Robots that fly ... and cooperate | Vijay Kumar

Prevoditelj: Recezent: Tilen Pigac - EFZG

Good morning.

Dobro jutro.

I'm here today to talk about autonomous flying beach balls.

Danas ću govoriti

o samoupravljajućim, letećim... loptama za plažu.

(Laughter)

Ma ne, o okretnim lebdećim robotima poput ovog.

No, agile aerial robots like this one.

I'd like to tell you a little bit about the challenges in building these,

Nekoliko riječi o izazovima pri građenju robota

kao i mogućnostima koje pruža

and some of the terrific opportunities for applying this technology.

primjena ove tehnologije.

So these robots are related to unmanned aerial vehicles.

Dakle, ovi roboti

su svojevrsne bezpilotne letjelice.

Međutim ovdje prikazane letjelice su velike.

However, the vehicles you see here are big.

Budući da teže stotine kilograma、

They weigh thousands of pounds, are not by any means agile.

pokretljivost im je otežana.

They're not even autonomous.

Štoviše nisu niti samoupravljajuće.

In fact, many of these vehicles are operated by flight crews

U stvarnosti, mnoge od ovih letjelica

su upravljane stručnim osobljem

poput grupe pilota

that can include multiple pilots,

operators of sensors,

kontrolorima senzora

and mission coordinators.

i zapovjednicima misije.

What we're interested in is developing robots like this --

Mi smo se pozabavili razvojem robota poput ovih --

prikazana su dva primjera

and here are two other pictures --

robota koji se mogu kupiti u trgovini.

of robots that you can buy off the shelf.

Dakle to su helikopteri sa četiri rotora-propelera

So these are helicopters with four rotors,

otprilike promjera jednog metra

and they're roughly a meter or so in scale,

i teže koji kilogram.

and weigh several pounds.

Opremili smo ih sa senzorima i procesorima,

And so we retrofit these with sensors and processors,

da bi mogli letjeti u zatvorenom prostoru

and these robots can fly indoors.

bez GPSa.

Without GPS.

Robot kojeg držim

The robot I'm holding in my hand

...ovaj...

is this one,

su kreirala dva studenta.

and it's been created by two students,

Alex i Daniel.

Alex and Daniel.

Težak je jedva

So this weighs a little more than a tenth of a pound.

40-tak grama

It consumes about 15 watts of power.

i pokreće ga snaga od 15 wati.

Kao što možete ocijeniti

And as you can see, it's about eight inches in diameter.

u promjeru je oko 20 cm.

Dopustite da ukratko pojasnim

So let me give you just a very quick tutorial

on how these robots work.

na koji način ovi roboti rade.

Robot ima četiri propelera.

So it has four rotors.

If you spin these rotors at the same speed,

Ako se svi propeleri okreću istom brzinom

the robot hovers.

robot lebdi u zraku.

If you increase the speed of each of these rotors,

Kada povećamo okretaje propelera

robot se uzdiže.

then the robot flies up, it accelerates up.

Of course, if the robot were tilted,

Ako je robot nagnut

prema horizontali

inclined to the horizontal,

then it would accelerate in this direction.

onda će se kretati u tom smjeru.

So to get it to tilt,

Evo jednog od dva načina kako nakrenuti robota.

there's one of two ways of doing it.

Na ovoj slici

So in this picture, you see that rotor four is spinning faster

uočite kako se 4. propeler okreće brže,

and rotor two is spinning slower.

a 2. propeler se okreće sporije.

And when that happens,

U tom trenutku

there's a moment that causes this robot to roll.

robot se nakrene.

I drugi način je

And the other way around,

if you increase the speed of rotor three and decrease the speed of rotor one,

da povećate brzinu 3. propelera

i smanjite brzinu 1. propelera,

čime je robot usmjeren prema naprijed.

then the robot pitches forward.

And then finally,

Konačno

if you spin opposite pairs of rotors

ako ubrzate dva

faster than the other pair,

nasuprotna propelera

then the robot yaws about the vertical axis.

robot će se okretati oko vertikalne osi.

U osnovi, ugrađeni procesor

So an on-board processor

essentially looks at what motions need to be executed

opaža koje kretnje trebaju biti izvršene,

usklađuje kretnje

and combines these motions,

te šalje primjerenu naredbu motorima

and figures out what commands to send to the motors --

600 puta u sekundi.

600 times a second.

That's basically how this thing operates.

Tako u osnovi robot radi.

Prednost ovakvog dizajna

So one of the advantages of this design

je, da s manjim komponentama

is when you scale things down,

robot postaje okretniji.

the robot naturally becomes agile.

Na ovoj slici R predstavlja

So here, R is the characteristic length of the robot.

karakterističnu duljinu robota.

U osnovi to je radius robota.

It's actually half the diameter.

And there are lots of physical parameters that change as you reduce R.

Mnogi fizikalni parametri se mijenjaju

kako smanjujemo R.

The one that's most important is the inertia,

Najvažniji parametar

je tromost, ili otpor kretanju.

or the resistance to motion.

Ispada da je

So it turns out the inertia, which governs angular motion,

tromost, o kojoj ovisi kretanje pod kutom,

proporcionalna R na petu potenciju.

scales as a fifth power of R.

Sa smanjivanjem R

So the smaller you make R,

the more dramatically the inertia reduces.

tromost dramatično pada.

So as a result, the angular acceleration,

To rezultira kutnom akceleracijom

denoted by the Greek letter alpha here,

označenom grčkim slovom alfa

goes as 1 over R.

približno 1/R.

It's inversely proportional to R.

Dakle, obrnuto proporcionalna s R.

The smaller you make it, the more quickly you can turn.

Što je robot manji, to je okretniji.

Što je očito iz ovih snimki.

So this should be clear in these videos.

Promotrite robota, desno dolje,

On the bottom right, you see a robot performing a 360-degree flip

kako se okrenuo 360 stupnjeva.

in less than half a second.

u djeliću sekunde.

Multiple flips, a little more time.

Višestruki okreti - malo više vremena.

Ugrađeni procesori

So here the processes on board

zaprimaju informaciju od ugrađenih

are getting feedback from accelerometers and gyros on board,

akcelerometara i žiroskopa

and calculating, like I said before,

te procesiraju, kao što rekoh,

commands at 600 times a second,

te informacije 600 puta u sekundi

to stabilize this robot.

stabilizirajući robota.

So on the left, you see Daniel throwing this robot up into the air,

Na dnu lijevo, uočite Daniela kako baca robota u zrak

and it shows you how robust the control is.

demonstrirajući koliko je robot robustan.

Bez obzira kako ga bacite

No matter how you throw it,

the robot recovers and comes back to him.

robot se dočeka na noge i doleti nazad.

So why build robots like this?

Koja je svrha graditi ovakve robote?

Može biti višestruka.

Well, robots like this have many applications.

Robote možete poslati u zgrade pod ugrozom

You can send them inside buildings like this,

as first responders to look for intruders,

primjerice kao izvidnicu za otkrivanje provalnika,

otkrivanje biokemijskog hazarda,

maybe look for biochemical leaks,

ili curenja plina.

gaseous leaks.

Možete ih koristiti i kao

You can also use them for applications like construction.

programirane graditelje.

So here are robots carrying beams, columns

Uočite kako roboti nose gredice i stupiće

i sklapaju ih u kockaste konstrukcije.

and assembling cube-like structures.

I'll tell you a little bit more about this.

O tome malo više kasnije.

The robots can be used for transporting cargo.

Ove robote možemo koristiti i za prijenos tereta.

So one of the problems with these small robots

Međutim, s obzirom na njihovu veličinu

is their payload-carrying capacity.

kapacitet nosivosti je malen.

So you might want to have multiple robots carry payloads.

Stoga bi bilo dobro udružiti robote

za prijenos tereta.

This is a picture of a recent experiment we did --

Evo snimke nedavnog pokusa --

doduše već ima tome --

actually not so recent anymore --

in Sendai, shortly after the earthquake.

u Sendaiu, netom iza potresa.

So robots like this could be sent into collapsed buildings,

Roboti mogu biti poslani u ruševine

kako bi se procjenila šteta

to assess the damage after natural disasters,

ili u objekte s reaktorima

or sent into reactor buildings,

to map radiation levels.

da izmjere razinu radijacije.

So one fundamental problem that the robots have to solve

Osnovni problem

s kojim se roboti moraju nositi, samoinicijativno

if they are to be autonomous,

is essentially figuring out how to get from point A to point B.

je u biti zaključivanje

kako doći od točke A do točke B.

So this gets a little challenging,

Dakle stvari postaju složenije

because the dynamics of this robot are quite complicated.

s porašću dinamike robota.

In fact, they live in a 12-dimensional space.

U načelu, roboti žive u 12-dimenzionalnom prostoru.

Pa smo morali tome doskočiti.

So we use a little trick.

We take this curved 12-dimensional space,

Uzeli smo zakrivljeni 12-dimenzionalni prostor

i transformirali ga

and transform it into a flat, four-dimensional space.

u ravninski 4-dimenzionalni prostor.

Ovaj 4-dimenzionalni prostor

And that four-dimensional space consists of X, Y, Z,

se sastoji od X, Y, Z i osnog kuta.

and then the yaw angle.

And so what the robot does,

Što je robotu činiti

is it plans what we call a minimum-snap trajectory.

je isplanirati putanju najmanje akceleracije.

Da se prisjetimo fizike,

So to remind you of physics:

You have position, derivative, velocity;

imamo položaj, te 1. derivaciju položaja - brzinu,

then acceleration;

zatim akceleraciju - 2. vremenska derivacija položaja (d2x/dt2)

and then comes jerk,

impuls - 3. derivacija položaja (d3x/dt3)

and then comes snap.

'cimanje' - 4. derivacija položaja (d4x/dt4).

So this robot minimizes snap.

Robot minimizira cimanje.

So what that effectively does,

Rezultat toga je

is produce a smooth and graceful motion.

glatko, ravnomjerno kretanje robota.

And it does that avoiding obstacles.

I to pri zaobilaženju zapreka.

Dakle, putanje najmanje akceleracije u ovom ravninskom prostoru

So these minimum-snap trajectories in this flat space are then transformed

potom transformiramo nazad

back into this complicated 12-dimensional space,

u složeni 12-dimenzionalni prostor,

koje robot mora slijediti

which the robot must do for control and then execution.

radi provjere i potom izvršavanja.

So let me show you some examples

Evo nekoliko primjera

of what these minimum-snap trajectories look like.

kako putanje najmanje akceleracije izgledaju.

U prvoj snimci

And in the first video,

you'll see the robot going from point A to point B,

vidjet ćete kretanje robota od točke A do točke B

ali kroz zadanu točku.

through an intermediate point.

(Whirring noise)

Robot je sposoban

So the robot is obviously capable of executing any curve trajectory.

pratiti zakrivljene putanje.

Ovo su kružne putanje

So these are circular trajectories,

where the robot pulls about two G's.

na kojima robot potegne 2G.

Here you have overhead motion capture cameras on the top

Ovdje imamo stropne kamere

that tell the robot where it is 100 times a second.

koje signaliziraju robotu njegov položaj 100 puta u sekundi.

It also tells the robot where these obstacles are.

Također daju robotu do znanja gdje su prepreke.

Te prepreke mogu biti pokretne.

And the obstacles can be moving.

And here, you'll see Daniel throw this hoop into the air,

Ovdje vidite Daniela kako baca kolut u zrak,

while the robot is calculating the position of the hoop,

a robot računa položaj koluta

pokušavajući odrediti pravi trenutak za prolazak kroz kolut.

and trying to figure out how to best go through the hoop.

Kao znanstvenici, primorani smo izvoditi

So as an academic,

we're always trained to be able to jump through hoops

akrobacije kako bi fondovi za naše projekte bili odobreni,

to raise funding for our labs,

pa smo odlučili i robote naučiti akrobacijama.

and we get our robots to do that.

(Applause)

(Pljesak)

Dodatna stvar koju robot može učiniti

So another thing the robot can do

je memoriranje dijelova putanje

is it remembers pieces of trajectory

that it learns or is pre-programmed.

koju je savladao ili mu je predprogramirana.

Ovdje vidite robota

So here, you see the robot combining a motion that builds up momentum,

kako kombinira kretnje

i vreba pravi trenutak

and then changes its orientation and then recovers.

u kojem će se nagnuti i potom vratiti u početni položaj.

Na to je prisiljen jer svjetli otvor prozora

So it has to do this because this gap in the window

is only slightly larger than the width of the robot.

je tijesan za njegovu širinu.

Poput skakača s daske

So just like a diver stands on a springboard

and then jumps off it to gain momentum,

kada odskoči da bi dobio zamah,

potreban za piruetu, i dvostruki salto

and then does this pirouette, this two and a half somersault through

da bi se na kraju ponovo ispružio,

and then gracefully recovers,

this robot is basically doing that.

ovaj robot u osnovi čini isto.

So it knows how to combine little bits and pieces of trajectories

Dakle zna kako usuglasiti više manjih putanja

to do these fairly difficult tasks.

u jednu znatno složeniju.

Promjenimo temu.

So I want change gears.

So one of the disadvantages of these small robots is its size.

Jedan od nedostataka robota je veličina.

I kao što rekoh

And I told you earlier

that we may want to employ lots and lots of robots

trebat ćemo upregnuti veliki broj robota

da bi nadomjestili taj nedostatak.

to overcome the limitations of size.

Prva poteškoća je

So one difficulty is:

How do you coordinate lots of these robots?

kako koordinirati jato robota?

And so here, we looked to nature.

Okrenuli smo se majci prirodi.

So I want to show you a clip of Aphaenogaster desert ants,

Pogledajmo slijedeći prilog

s pustinjskim mravima Aphaenogaster.

kako nose teret, u laboratoriju profesora Stephen Pratta.

in Professor Stephen Pratt's lab, carrying an object.

Ovo je komadić smokve.

So this is actually a piece of fig.

Ako im podmetnete bilo koji predmet preliven smokvinim sokom

Actually you take any object coated with fig juice,

mravi će ga odvući u mravinjak.

and the ants will carry it back to the nest.

So these ants don't have any central coordinator.

Mravi očito nemaju koordinatora.

Ali osjećaju ostale iz skupine.

They sense their neighbors.

Ne postoji izravna komunikacija.

There's no explicit communication.

Ali kako osjećaju ostale u skupini

But because they sense the neighbors

i kako osjećaju predmet,

and because they sense the object,

they have implicit coordination across the group.

koordinacija je ostvarena posredno.

Takvu sličnu koordinaciju

So this is the kind of coordination we want our robots to have.

smo željeli ostvariti kod robota.

So when we have a robot which is surrounded by neighbors --

Ako imamo robota

grupiranog u jato

promotrimo robota I i robota J --

and let's look at robot I and robot J --

želimo robote osvijestiti

what we want the robots to do,

da paze na međusobni razmak

is to monitor the separation between them,

as they fly in formation.

kada tvore formaciju.

And then you want to make sure

Želimo da se taj razmak

that this separation is within acceptable levels.

održava unutar određene tolerancije.

Da ponovimo, roboti prate ovu toleranciju

So again, the robots monitor this error

and calculate the control commands 100 times a second,

i proračunavaju kontrolne naredbe

100 puta u sekundi,

which then translates into motor commands,

što znači da motor zaprima 600 naredbi u sekundi.

600 times a second.

I to treba biti učinjeno

So this also has to be done in a decentralized way.

decentralizirano.

Again, if you have lots and lots of robots,

Ako imate jato robota,

bilo bi nemoguće koordinirati ih centralno

it's impossible to coordinate all this information centrally

fast enough in order for the robots to accomplish the task.

zadovoljavajućom brzinom za ostvarenje zadaće.

Plus, the robots have to base their actions only on local information --

Dodatno, roboti moraju zasnivati kretnje

isključivo na lokalnim informacijama

koje zaprimaju od susjednih robota.

what they sense from their neighbors.

I napokon

And then finally,

we insist that the robots be agnostic to who their neighbors are.

inzistirali smo da su roboti agnostični

prema susjedima.

So this is what we call anonymity.

To smo nazvali anonimnost.

Sada ću vam pokazati

So what I want to show you next is a video of 20 of these little robots,

snimak

sa 20 robota

flying in formation.

koji lete u formaciji.

Svaki robot promatra položaj susjeda.

They're monitoring their neighbors' positions.

I tako održavaju formaciju.

They're maintaining formation.

The formations can change.

Formacija se može mijenjati.

They can be planar formations,

Može biti ravninska

they can be three-dimensional formations.

ili prostorna.

Kao što ovdje vidite,

As you can see here,

they collapse from a three-dimensional formation into planar formation.

roboti iz prostorne formacije tvore ravninsku.

I da bi letjeli kroz prepreke

And to fly through obstacles,

they can adapt the formations on the fly.

sposobni su prilagoditi formaciju u letu.

Roboti mogu prići blizu jedan drugom.

So again, these robots come really close together.

As you can see in this figure-eight flight,

Kao što vidite u ovoj formaciji u obliku osmice,

they come within inches of each other.

roboti prilaze jedni drugima unutar par centimetara.

I unatoč aerodinamičnom utjecaju

And despite the aerodynamic interactions with these propeller blades,

propelerskih elisa

they're able to maintain stable flight.

održavaju stabilan let.

(Applause)

(Pljesak)

Kada ih jednom naučite letjeti u formaciji

So once you know how to fly in formation,

možete podići predmete surađujući.

you can actually pick up objects cooperatively.

To ujedno znači

So this just shows that we can double, triple, quadruple

da možemo udvostručiti, utrostručiti ili učetverostručiti

snagu robota

the robots' strength,

by just getting them to team with neighbors, as you can see here.

usklađujući njihovo djelovanje, kao što je prikazano.

Jedan od nedostataka ovakvog pristupa

One of the disadvantages of doing that is, as you scale things up --

je, kako teret postaje teži...

so if you have lots of robots carrying the same thing,

i ako uposlite jato robota za prijenos tereta

u osnovi povećavate tromost,

you're essentially increasing the inertia,

and therefore you pay a price; they're not as agile.

čime naravno jato postaje manje okretno.

But you do gain in terms of payload-carrying capacity.

Ali tako možete prenašati teže objekte.

Još jedna primjena koju želim pokazati --

Another application I want to show you -- again, this is in our lab.

ponovo, u našem laboratoriju.

This is work done by Quentin Lindsey, who's a graduate student.

Ovo je napravio naš postdiplomac Quentin Lindsey.

Njegov algoritam u osnovi naređuje robotima

So his algorithm essentially tells these robots

kako da samostalno sklope

how to autonomously build cubic structures

prostorne rešetkaste konstrukcije

iz gredica.

from truss-like elements.

Njegov algoritam govori robotu

So his algorithm tells the robot what part to pick up,

koji dio podići,

te kada i gdje ga ugraditi.

when, and where to place it.

Na ovoj snimci možete vidjeti --

So in this video you see --

and it's sped up 10, 14 times --

ubrzano 10 - 14 puta --

tri različite konstrukcije sklapane robotima.

you see three different structures being built by these robots.

Roboti su samoupravljajući,

And again, everything is autonomous,

i sve što Quentin treba učiniti

and all Quentin has to do

is to give them a blueprint of the design that he wants to build.

je dostaviti im nacrt

konstrukcije koja treba biti sklopljena.

So all these experiments you've seen thus far,

Dakle, svi ovi pokusi koje ste vidjeli,

all these demonstrations,

sve ove demonstracije,

have been done with the help of motion-capture systems.

su ostvarene uz pomoć sustava za praćenje kretnji.

Međutim, što se događa kada napustite laboratorij

So what happens when you leave your lab,

i otisnete se u stvaran svijet?

and you go outside into the real world?

And what if there's no GPS?

I što kada nema GPSa?

So this robot is actually equipped with a camera,

Primjerice, ovaj robot

je opremljen kamerom

i laserskim daljinomjerom, laserskim skenerom.

and a laser rangefinder, laser scanner.

I koristi ove senzore

And it uses these sensors to build a map of the environment.

kako bi izgradio kartu okruženja.

Takva karta sadrži objekte --

What that map consists of are features --

poput vrata, prozora,

like doorways, windows, people, furniture --

ljudi, namještaja --

i onda robot procjenjuje gdje se nalazi

and it then figures out where its position is,

u odnosu na te objekte.

with respect to the features.

So there is no global coordinate system.

Sve bez globalnog koordinatnog sustava (GPS).

Koordinatni sustav je definiran robotom,

The coordinate system is defined based on the robot,

where it is and what it's looking at.

njegovim položajem, i time što promatra.

And it navigates with respect to those features.

I navođenje se odvija u odnosu na te objekte.

Evo isječka

So I want to show you a clip

of algorithms developed by Frank Shen and Professor Nathan Michael,

s algoritmima koje je razvio Frank Shen

i profesor Nathan Michael

that shows this robot entering a building for the very first time,

koji prikazuje robota kako ulazi u zgradu po prvi puta

and creating this map on the fly.

i stvara kartu u letu.

So the robot then figures out what the features are,

Robot razaznaje objekte i

it builds the map,

iscrtava kartu (model).

it figures out where it is with respect to the features,

Izračunavajući udaljenosti do objekata

and then estimates its position 100 times a second,

robot određuje svoj položaj

100 puta u sekundi

allowing us to use the control algorithms that I described to you earlier.

uz korištenje kontrolnog algoritma

koji sam opisao ranije.

So this robot is actually being commanded remotely by Frank,

Dakle ovaj robot zaprima

Frankove naredbe sa udaljene lokacije.

Ali robot također može

but the robot can also figure out where to go on its own.

samostalno procijeniti gdje treba ići.

Recimo da želim poslati robota u zgradu

So suppose I were to send this into a building,

za koju nemam predodžbu o unutrašnjosti.

and I had no idea what this building looked like.

Jednostavno pošaljem robota da

I can ask this robot to go in,

kreira kartu (model)

create a map,

i pozovem ga nazad da mi predoči što je zabilježio.

and then come back and tell me what the building looks like.

So here, the robot is not only solving the problem

Dakle, robot ne samo da rješava problem

kako stići od točke A do točke B na ovoj karti,

of how to go from point A to point B in this map,

but it's figuring out what the best point B is at every time.

nego i procjenjuje

koja je optimalna točka B.

U osnovi, robot zna da treba ići

So essentially it knows where to go

to look for places that have the least information,

prema lokacijama o kojima ima najmanje informacija.

I na taj način iscrtava kartu (model).

and that's how it populates this map.

So I want to leave you with one last application.

I da zaključim

s još jednom primjenom robota.

And there are many applications of this technology.

A postoje i mnoge druge primjene.

I'm a professor, and we're passionate about education.

Ja sam profesor, i kao takav strastven za edukaciju.

Ovakvi roboti mogu promijeniti

Robots like this can really change the way we do K-12 education.

način na koji se odvija nastava.

Ali s obzirom da smo u Južnoj Kaliforniji,

But we're in Southern California,

blizu Los Angelesa,

close to Los Angeles,

morat ću zaključiti

so I have to conclude with something focused on entertainment.

u holivudskom tonu.

Neka to bude glazbeni video.

I want to conclude with a music video.

Želim predstaviti Alexa i Daniela,

I want to introduce the creators, Alex and Daniel, who created this video.

kreatore videa.

(Pljesak)

(Applause)

So before I play this video,

No prije nego prikažemo video,

I want to tell you that they created it in the last three days,

želim napomenuti da su ga snimili u zadnja tri dana

after getting a call from Chris.

nakon poziva od Chrisa (TED kuratora).

And the robots that play in the video are completely autonomous.

I roboti koji sviraju na snimci

su potpuno samostalni.

You will see nine robots play six different instruments.

Vidjet ćete 9 robota kako svira 6 instrumenata.

Naravno, eksluzivno za TED 2012.

And of course, it's made exclusively for TED 2012.

Pogledajmo.

Let's watch.

(Sound of air escaping from valve)

(Music)

(Whirring sound)

(Music)

(Glazba)

(Pljesak)

(Applause) (Cheers)